The　production　efficiency　of　plastic　gears　is　high,　and　their　batches　are　large;　while　the　traditional　manual　inspection　is　low　and　the　accuracy　is　poor.　To　ensure　product　quality　and　production　efficiency,　a　rapid　sorting　and　inspection　system　for　plastic　gears　was　developed.　The　system　was　combined　with　a　production　line　and　used　a　conveyor　belt　for　loading,　a　glass　rotary　for　transmission,　and　a　pneumatic　device　for　sorting.　Plastic　gear　images　were　captured　using　an　industrial　camera,　and　basic　parameters　such　as　the　tip　diameter,　the　hole　diameter,　and　the　concentricity　were　measured　by　a　machine　vision　method;　the　tolerance　ranges　were　compared,　in　order　to　judge　the　product　quality.　By　using　algorithms　such　as　image　segmentation,　edge　extraction,　feature　extraction,　and　defect　identification,　surface　defects　such　as　missing　teeth,　black　spots,　and　burr　were　inspected.　Finally,　the　system　separated　good　and　defective　products　according　to　the　test　results.　The　system　could　carry　out　fast,　online,　automatic,　and　non-contact　separation　and　inspection.　The　test　results　showed　that　the　measurement　repeatability　of　the　system　was　less　than　0.005　mm,　the　inspection　speed　was　greater　than　150　piece/min,　the　defect　identification　and　sorting　functions　were　accurate,　and　the　system　could　run　stably　for　a　long　time.　Hence,　the　proposed　system　can　meet　the　requirements　of　large-scale,　high-efficiency,　and　high-reliability　sorting　and　inspection　at　a　plastic　gear　production　site.　©　2020,　Science　Press.　All　right　reserved.